Revisiting Two Old Bond Topics: The Government Bond Dealer and New-Seasoned Bond Spreads

I study the relationship between two old bond topics: The Government bond dealer and new-seasoned bond spreads. I establish empirical support for the Gaines (1962) 'amplification hypothesis' that changes in net positions have a causal effect on the spreads offered by primary dealers: net Treasury positioning is weakly exogenous for the AAA-10Y, BAA-AAA and 10Y spline errors. I also find support for the 'textbook' Scott (1965) hypothesis: that dealers adjust total net positions by changing spreads. I reconcile the two contradictory views by highlighting the distinction between total net positions which includes all fixed income securities and net positioning in longer duration Treasuries. An offshoot of the investigation is clarification into the role Treasury supply has on credit spreads.